Have an idea?

Visit Sawtooth Software Feedback to share your ideas on how we can improve our products.

Which is better, referring to a questionnaire or adding the questions to the original ?


I have a relatively long questionnaire (66 questions up to this point), measuring health state utilities, composed of several parts. Besides the demographics and general information collection, there are CBC, MaxDiff and Time Trade-off parts (TTO is a method where the respondent compares living in a certain ill health state for like 10 years to living in full health for a changing length of years). The TTO option was prepared via free formatting question (total of abouit 250+ questions but only around 15 of those is shown to each respondent) for another questionnaire that is relevant to the one I am doing. I have the following concerns:
1- Which option is better? adding the questions of that questionnaire to the original or referring to that questionnaire within the original questionnaire? knowing that we need to modify the original a little bit (I asked that in a different question) how would the referring option be managed?

2- We need to randomize the respondents for the order of seeing the 3 parts (TTO, MaxDiff & CBC), does that affect our decision in the first point? & how is this done in Lighthouse? (I have an idea that it should be through the randomization option & building lists but I did not get how exactly this could be done)

I wrote this question before but and it did not show up that time, I just found it in the asked questions but maybe it was not clear so I rephrased it.

Thank you very much and have a great day.

asked Apr 5, 2019 by AMYN Bronze (2,980 points)
I'm not well versed in the benefits and drawbacks of running multiple questionnaires instead of just one, but I may be able to offer some insight on implementations.  There are two common approaches to doing something like this:

First, you could add a predefined list to your study with three items in it, and then add a constructed list that randomizes that first list.  Then in Randomization, you can use randomized blocks to define three blocks in the code that will be shown in the same order as the items in that constructed list.  I feel like this is the cleanest option, but it's only possible with all questions in the same survey.

The other route would involve skips.  If you randomly assign each respondent a number one through six, you could add skips to the relevant parts of your survey to skip respondents around depending on which random value they received.  A bit more setup, but you might have to do something akin to this if you want to skip respondents around several surveys.

Your solution to the original question

Please only use this to answer the original question. Otherwise please use comments.
Your name to display (optional):
Privacy: Your email address will only be used for sending these notifications.
Anti-spam verification:

To avoid this verification in future, please log in or register.